Skullduggery Thread for Demohacking

[YossarianLives]

Welcome to the skullduggery thread for demohacking! This is where we will make educated guesses and calculations to determine what the other teams are doing. We will track technologies researched, military unit and improvement builds, new cities and total population, whips, and anything else we can infer from the demographics numbers, civstats, top 5 cities screen, espionage screen, etc.

INSTRUCTIONS TO ACCESS THE SPREADSHEET
1. Go to www.drive.google.com
2. Enter email diplo.civfanatics
3. Enter password hangingGardens

If you would like access to view the spreadsheet from your own personal Google account, just give me your gmail and I will add you to the list. The more teammembers that are viewing and updating the spreadsheet, the better, so I highly encourage this for everyone.

Edit:I set the spreadsheet to private and shared it with diplo.civfanatics@gmail.com. If anyone would like their email added to the list so you can view it without logging on to our CFC email account, just let me know and I'd be happy to add you (or you can add yourself by going through the CFC account).

Here are some useful links that explain how the numbers work in the art of demohacking :
Demographics screens explained
Scoring explained
Another Demos discussion
RB article on using the espionage screen to predict city builds
I’ll add these to the Resources and References thread.

ETA: RB's Espionage thread from the Poly Demogame, with some great formulas for determining how many EP are needed for graphs / tech visibility / Espy missions / etc.

 

[YossarianLives]

Score
Score is made up of four factors: Population, Land, Technologies and Wonders. We can see this on the very first turn by mousing over our score.

There are 5000 possible points for population. The game determines a theoretical “max population” based on the map, and then determines how many points you get per pop based on that. The equation is (5000/max pop)*current pop. See this example in the Scoring explained link. It is showing the score includes 2613 points from population (426/815). That is, (5000/815)*426=2613.49 rounded down. You can also see how much each pop is worth. In this example, (5000/815)*1=6.13 points per pop.

There are 2000 possible points for land. The game shows you how many total land tiles there are on the map, and how many of those you own. The equation is the same as for pop: (2000/total tiles)*owned tiles. Again, using the example from the same link, there are 1069 points from land (479/896). That is, (2000/896)*479=1069.19 rounded down. You can also see how much each land tile is worth. In this example, (2000/896)*1=2.23 points per land tile. Land tiles only count once they have been owned for 20 turns, so we should mark ahead when we expect land score increases so they don’t throw us off.

Technologies are worth 5.99*Tech Era Value. Ancient is worth 1, Classical is 2 and so on.

There are 1000 possible points for wonders. There are 62 wonders in the game, so we get 1000/62=16.13 points per wonder (including National wonders)

In the spreadsheet, we will record each civ’s score on each turn, taking careful note of all score changes (and when they happened). We will use the next two lines to record the causes for any score changes (ie, “discovered TW”, “pop increase”, “expanded borders”). Once we discover how much each pop and land is worth on the first turn, I will add that to the spreadsheet. More info on how we will determine what causes the score increases to follow.

 

[YossarianLives]

Population
So, the way the spreadsheet works is that we’re going to record how many cities of each size each team has. At the beginning, each team will have one size-1 city, then some will grow to size-2 and they will have zero size-1 cities and one size-2 city. There are equations to auto-calculate the number of cities, total world population, average city size, etc. We’ll know when each team plants a new city because it will be a score increase in the middle of the turn. The easiest way to spot pop increases (excluding new cities) is looking for corresponding score and power increases (end of turn score increase instead of middle of the turn increase, and 1k power for every 2 pop). We will also keep an eye on the max, min and ave pop numbers in the demos screen. If we know how many pop points a team has, we should be able to calculate the size of each city.

Ok, I’m actually going out to a barbecue right now. If I get a chance to finish posting about the last two tabs tonight, I will. Otherwise, it will probably be in about 24 hours before I’ll get to finish.

Happy 4th of July to all the Americans here!

 

[Bowsling]

Set the spreadsheet to private and put it in the team account. Everyone is allowed to access it. The link doesn't work for me anyway.

EDIT: Removed the extra characters from the URL and loaded it. Nice work!

 

[YossarianLives]

Tech
Ok, back from the barbecue. A little more tipsy than before, but let’s see if I can finish out the explanation of the spreadsheet. Has anybody been able to take a look at it, or is the link completely broken?

We should always be able to tell when a team researches a tech because we know exactly what their score increase will be (6 pts for an ancient tech, 12 pts for a classical tech, etc.) The trick is figuring out which tech they got. We will need to look at how long it took them between techs (and compare that to their GNP number), whether they got a power increase with the tech, whether a religion was founded, and what makes sense logically for them to tech next.

For example, we know RB starts with Agri and TW, so if they get a tech without a power increase 10 or 11 turns after founding, we can logically deduce that they probably teched pottery (they could tech hunting or myst faster than that, mining or hunting would come with a power increase, and it just makes sense for them to open with pottery).

We may not always be 100% sure that we have correctly predicted the researched tech, but we should be able to make some pretty educated guesses most of the time. Over time, keeping track of how many techs each team has researched will also give us a pretty decent picture of a team’s overall research power. Once we get alphabet, we will be able to see other team’s techs more directly, so that should make updating the spreadsheet easier. The costs on the spreadsheet are from this page, Civilization 4: Technologies, but they’re not what come up when I actually open up the game, so we might have to fix that.

 

[YossarianLives]

Power
Power is another thing that is important to keep track of so we know which teams are threats to us, but can be very difficult to get 100% right all of the time. Power is calculated using four components; population, technologies, city improvements/wonders and units. Population is always 1k power for every two pop. The power value of techs, improvements and units vary. The spreadsheet contains all the values for these factors. At the beginning, we’ll have to watch out for warriors, galleys (2k each), archers (3k), spearmen, chariots (4k) and axemen (6k).

We will always know the value of our own power, our opponent’s top power, lowest power, and the average of the middle six opponents (the formula for that is (Rival Average*8-(Rival Best + Rival Worst))/6). We will have to make some educated guesses to determine who is who, and also some assumptions based on the team’s decisions so far. For example, Spanish Poly and WPC will probably open with larger power scores than other teams.

Now that I’m thinking about it, I want to add another tab where we can just record the demo numbers from each turn, and a space for us to write notes with our explanations for the changes from one turn to the next. I’ll post again when that has been added.

I think there was talk about having this information in Dropbox as part of the app that’s being developed. Is there a way to do that with the Googledoc I made, or do we need to transfer the info to Dropbox somehow? I am very clueless on this type of thing, so please just let me know what needs to be done, if anything, to make this spreadsheet compatible with the app.

 

[YossarianLives]

Set the spreadsheet to private and put it in the team account. Everyone is allowed to access it. The link doesn't work for me anyway.

EDIT: Removed the extra characters from the URL and loaded it. Nice work!

Thanks Bowlsing. I think I followed your advise. I set the spreadsheet to private and shared it with our Google account (diplo.civfanatics). Is this what's easiest (and most secure) for everyone?

 

[YossarianLives]

Demographics

I finally added the Demographics tab to the spreadsheet. It actually doesn't contribute too much we won't get from the screenshots, except it will auto-calculate the rival average excluding rival best & worst, to get us a slightly more accurate number for the middle six rivals. We can also keep notes on the changes to the demos, such as which team we think went up or down in each category. I'm not including Approval Rate, Imports/Exports or Espionage numbers because I don't think there's as much to analyze about those.

I'm reposting this link because it gives a lot of great insight into the demos screen.

The GNP number is composed of commerce, culture and espionage. The equation for getting this number is total commerce+total culture+total espionage+(total research*research bonus)-expenses. The research bonus is (30%*# of known civs with tech/total opponents in game)+20% pre-req bonus(es). Civfr has the only cultural leader, so they will get an automatic +2 GNP per city they have. At the beginning everyone else should have a similar GNP, as the only things that will increase it is working a tile with 1 commerce, and researching a tech with a pre-req (or two).

The MFG number is more straight-forward. It is all production produced, taking modifiers from city improvements into account. A team focusing on high MFG production might need to be watched in case they are amassing an army with all those hammers.

The Crop Yield number is simply total food produced. Because food can be translated into production so efficiently through whips, a high Crop Yield is a very good predictor of how well teams are doing overall.

The Soldiers number takes a whole lot into account, from techs to improvements to population to actual unit count. The above post on Power goes into this in more detail.

Land Area is the number of land tiles*1000. This will let us know, for example, if teams have water tiles in their capital, confirm new cities and border pops, etc.

Population can help us determine the size of each city a team has. This is because population points increase exponentially as a city grows larger:

Size --> population value
1 --> 1k
2 --> 6k
3 --> 21k
4 --> 45k
5 --> 90k
6 --> 150k
7 --> 232k

So if a team has 3 cities and 33k pop points, then we know the cities must be size 3,2,2 since this is the only 3-city combo that equals 33k points (21k+6k+6k).

This all includes a lot of guesswork, and becomes much easier if we can unfog city tiles. If we can see the city tile, we can just look at the map to track improvements, and use the espionage screen to get even more information about what our opponents are doing.

 

[Caledorn]

This is impressive work! I realize I could log in and do this myself, but I'd like to ask before doing such a thing. Can we share this googledoc to our respective google accounts, so that we don't have to log in to the team account to look at it? If that is okay, then my gmail account is my nickname here at CFC followed by the at gmail.com (I won't touch it, as this is definitely not my department, but I would love to be able to track changes made to it - so if you add me, read-only access will suffice for my use)

 

[YossarianLives]

Yes, I have added your email account, and anyone else who would like to do so please feel free through the diplo.civfanatics account. Everyone can have editing privileges. I have the template on my computer that we can revert back to if anything gets irreparably changed, but I'm sure it'll be fine.

And we'll need lots of help trying to figure out who got what increases/decreases, and what they all mean, so the more people who are taking a look at the spreadsheet and chiming in, the better.

 

[Caledorn]

Thank you, Yossarian! I am impressed with the amount of work you've put into this spreadsheet.

 

[YossarianLives]

I'm very excited the game has started, and the skullduggery can begin!

When someone logs in, could we please get a screenshot of the mouseover of our score. This will allow me to complete the spreadsheet with the number of points per pop and land tile.

I realize this is a much more difficult request to fulfill, but if we can get screenshots of the demo screen after each other team has founded their city, that will eliminate a lot of guesswork and allow us to get a very accurate picture of the GNP, MFG and crop yields of each team.

 

[YossarianLives]

Thank you, Yossarian! I am impressed with the amount of work you've put into this spreadsheet.

Credit has to go to Cyneheard of Realms Beyond, as the spreadsheet and equations are very much based on the one he is using in the ongoing Realms Beyond Pitboss 5. He is a true master of this art and will certainly be outperforming me in demohacking (Cloak and Dagger by RB terminology) in this game. All the more reason I'll need lots of help here!

 

[2metraninja]

First, I want to say hip-hip-hooray! for Yossarian for taking this hard work at heart.

I'll have to read all this to comment more in-depth, but one things I wonder - how we are supposed to reverse-engineer border pop-ups when there are 20 turns delay between them showing up on the score?

 

[YossarianLives]

There is a 20 turn delay for it to show up on the score, but it will appear on the demos screen immediately. Plus, we should be able to predict exactly when the capitals pop borders for the first part of the game, and if we can track who founds religions, who gets wonders, etc. through score increases, then we should be able to predict border pops a little further into the game. As more cities are founded and more factors come into play, it gets harder and harder, so hopefully by then we have better map info so we don't have to rely on reading the demos.

We'll have to make notes on the spreadsheet so we can remember to take border pops into account when we're trying to figure out score increases.

 

[Bowsling]

Team Apolyton has finished their turn without building their city!

 

[2metraninja]

Team Apolyton has finished their turn without building their city!

This is not so unusual, what is unusual is the speed they did their turn. Can mean two things:
1/ They have discussed everything in advance and they know very well exactly what they are doing.
2/ The one playing their turn dont have a clue what to do and just play hoping for the best

 

[Bowsling]

Yes, I suspect they are going for the plains hill location (assuming of course that we all have similar starts, which may be wrong).

 

[YossarianLives]

Wow, yes, civstats is showing a lot to analyze. I don't have a lot of time right now, but I will definitely be taking a closer look later today. A couple things I noticed right away is that Apolyton logged in immediately once the option was available, then played, ended turn and logged out in three minutes! They definitely did not take any screenshots back to the team for discussion before making there decision not to settle. Also, RB settled their city this turn and have ended turn as well.

 

[YossarianLives]

I have updated the spreadsheet with the limited info available so far (RB founded a city and ended turn, score up to 29, Poly and CivPlayers ended turn without founding a city, score remains 11).

Apolyton's actions are most surprising. They obviously knew they were NOT going to settle on the first turn, and no scouting was going to change that. If we assume that we have mirrored starting positions, I would have to guess that means they are moving for the PH where the warrior started. However, I find it odd that they would choose a leader with two economic traits, and then move away from the commerce-rich FP's. They start with Agri and TW, so they could have gone straight to Pottery for early FP cottages. Therefore, this lends evidence to the fact that starts are not mirrored.

I am more hesitant to make any assumptions about why CivPlayers might have moved. They also played very, very quickly, ending turn one minute after logging in for the first time. Later they logged back in for a short ten minutes (perhaps taking screenshots?)

CivFr and the Germans have been in and out of the game all day, but have not ended turn. They are definitely having more discussion about what to do.

Well, that's it for now. I see some concern in the public threads that the aptmod website reveals information that we should not have access to. If this is the case and is going to be happening throughout the game, should we take advantage of the info? I have no idea what info is revealed and if it will help with demohacking. Does anyone have any input on this? Should I register to see the site?

Edit: If we can see the demographics screen with only RB having founded a city, we should be able to tell, for example, if they are working a tile with commerce, and if they founded on a PH for the extra hammer. Of course, with all the logging in and out they are doing, I believe they can change what tile they work each time to throw off our attempts to decipher what they are doing (and I would not put it past them to do this).